A method for detecting the stress of the edge of an automobile glass on line by reflection
Patent Information
- Application Number
- CN202610715514.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-05-22
- Publication Date
- 2026-10-27
AI Technical Summary
[0003]随着汽车玻璃在交通等领域的应用日益广泛，市场对其生产效率和产品质量的要求不断提升，现有的离线和抽检方式存在检测滞后性，无法及时反馈生产工艺缺陷，影响生产调整效率，已难以满足工业化连续生产的需求
[0015]本发明的有益效果是：通过创新采用反射式检测原理，以传送带或黑色陶磁镀层的漫反射光作为背景光并保留偏振特性，配合像素偏振相机的四方向偏振图像采集与相位差计算，有效突破了传统透射式检测无法穿透汽车玻璃边缘黑色陶瓷涂层的局限。借助图像相减及增强处理技术，成功抵消背景噪声并放大应力相关前景信息，结合图像识别算法，实现了玻璃磨边区光弹条纹的快速精准识别，检测结果与实际应力线性相关，确保了边缘最大应力检测的准确性，完全满足国标采样点的检测要求。
Smart Images
Figure CN122982467A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of online automotive glass testing technology, and particularly relates to a method for reflective online testing of automotive glass edge stress. Background Technology
[0002] Edge stress in automotive glass is a core indicator determining its mechanical properties, safety stability, and service life. National standards have clear quantitative requirements for this indicator, making accurate detection of edge stress during online production crucial. Traditional automotive glass stress testing often employs offline sampling or direct transmission online testing. Offline testing is inefficient, cannot achieve full batch coverage, and easily leads to substandard products entering the market. Direct transmission testing is impractical because automotive glass often has a black ceramic coating that prevents light from penetrating. Furthermore, traditional reflective testing is even less efficient due to the extremely weak light reflection from the black ceramic coating.
[0003] With the increasingly widespread application of automotive glass in transportation and other fields, the market demands for production efficiency and product quality are constantly rising. Existing offline and sampling inspection methods suffer from detection lag, failing to provide timely feedback on production process defects and impacting production adjustment efficiency, making them unsuitable for the needs of continuous industrial production. Therefore, there is an urgent need for a highly efficient, accurate, and non-disruptive reflective online inspection technology to solve these problems. Summary of the Invention
[0004] The purpose of this invention is to address the aforementioned technical problems by providing a method for reflective online detection of edge stress in automotive glass.
[0005] In view of this, the present invention provides a method for reflective online detection of edge stress in automotive glass, comprising the following steps: A pixel polarization camera is fixed above the conveyor belt. The camera is equipped with a telecentric lens. The telecentric lens has a depth of field that is not less than the thickness of the glass and is precisely focused on the edge of the glass. The lens magnification ratio is adapted to the size of the camera target surface, so that the imaging range covers the unidirectional stress area of the edge of the car glass and a margin is reserved for glass position deviation. An integrated light source is set in the camera optical path. The integrated light source is a monochrome LED surface light source with a hole in the middle. A circular polarizer with a polymer material substrate is laid on its light-emitting surface. The polarizer has a hole in the middle with the same size as the central hole of the light source and is fixed by pressing with a glass plate. The camera lens takes pictures of the edge of the glass through the central hole of the light source. A sensing sensor is set in front of the glass traveling on the conveyor belt. When the sensing sensor detects the arrival of the glass, it emits an electrical pulse to trigger the pixel polarization camera to take a picture. The sensing sensor determines the presence of the object by emitting and receiving light beams. The distance between the camera and the sensor is adjusted according to the size of the glass to ensure that the shooting position is above the sampling point specified by the national standard at the edge of the glass. After capturing images with a pixel polarization camera, pixels with polarization directions of 0°, 45°, 90°, and 135° are extracted from each group of 2×2 pixels to form I0 and I... 45 I 90 I 135 Four sub-images; the linear polarization degree (DoLP) field is calculated from the four sub-images using the following formula: DoLP=|Sin δ|=[(I0-I 90 ) 2 +(I 45 -I 135 ) 2 ] 0.5 / (I0+I 90 (2) The calculation results include and amplify stress-related foreground information. The order and center position of the photoelastic fringes in the glass edge area and nearby are identified by the image recognition algorithm. The phase difference of the glass edge is obtained by fitting and extrapolation. The stress can be obtained by multiplying it by the glass photoelastic fringes value. The entire testing process takes no more than 2 seconds, which is less than the time interval between two adjacent pieces of glass on the conveyor belt.
[0006] Preferably, the imaging range of the telecentric lens is 30mm.
[0007] Preferably, the trigger response time between the sensing sensor and the camera is in the microsecond range.
[0008] Preferably, each pixel of the pixel polarization camera is provided with a polarizer made of a nano grating in front of it, and every 2×2 pixels form a group, with the polarization directions of the pixels in the group being 0°, 45°, 90°, and 135° respectively.
[0009] Preferably, the aperture of the polarizer in the integrated light source is the same as the aperture of the central aperture of the light source, and it is fixed to the light-emitting surface by pressing with a glass plate.
[0010] Preferably, the phase difference δ at the outermost edge of the glass is converted into the actual stress value of the automotive glass by multiplying it by the photoelastic fringe value.
[0011] Preferably, formula (2) is used for I0 and I 90 I 45 I 135 The four sub-images were processed to eliminate identical background areas while preserving and amplifying stress-related foreground variation information.
[0012] Preferably, when using image recognition and artificial intelligence algorithms, the identification of the photoelastic fringes order and center position in the glass edge area and surrounding area is completed within 2 seconds. The correct distribution calculation of the phase difference field is also performed.
[0013] Preferably, the installation position of the sensing sensor and the distance between the sensor and the camera are dynamically adjusted according to the size of the glass of the car to be tested, so as to ensure that the shooting position is always located at the test point on the edge of the glass as specified by national standards.
[0014] Preferably, the polarized light emitted by the integrated light source is transmitted through the glass and forms diffuse reflected light that retains its polarization properties on the conveyor belt. This diffuse reflected light, as background light, passes through the glass again and carries stress information to be captured by the camera.
[0015] The beneficial effects of this invention are as follows: By innovatively adopting a reflective detection principle, using diffuse reflected light from a conveyor belt or black ceramic coating as background light while preserving polarization characteristics, and combining this with four-directional polarization image acquisition and phase difference calculation by a pixel polarization camera, it effectively overcomes the limitation of traditional transmission-based detection methods that cannot penetrate the black ceramic coating at the edge of automotive glass. Through image subtraction and enhancement processing techniques, background noise is successfully canceled and stress-related foreground information is amplified. Combined with image recognition algorithms, rapid and accurate identification of photoelastic fringes in the glass edge area is achieved. The detection results are linearly correlated with the actual stress, ensuring the accuracy of maximum edge stress detection and fully meeting the national standard sampling point detection requirements.
[0016] This invention's entire inspection process takes only about 2 seconds, less than the time interval between adjacent glass pieces on the production line. It allows for online inspection of the entire batch without interrupting the production process, completely solving the problems of low efficiency and incomplete coverage associated with offline inspection. Furthermore, the distance between the camera and sensor in the inspection system can be flexibly adjusted according to the glass size. The telecentric lens design is compatible with glass position deviations, and the equipment layout is simple and highly adaptable. This not only reduces the cost of industrial applications but also provides timely feedback on production process defects, offering real-time data support for production adjustments and significantly improving the quality control level and production efficiency of automotive glass production. Attached Figure Description
[0017] Figure 1 This is a schematic diagram of the online stress testing device of the present invention. Figure 2 This illustrates the light propagation path in the reflective detection method of this invention. Figure 3 The processing effect of the online stress detection of reflective automotive glass according to the present invention; (a) The original polarization image obtained by reflective detection; (b) Stress information diagram after processing / calculation using formula (1); Figure 4 This is a flowchart illustrating the determination of the phase difference at the outermost edge of the glass in this invention. Detailed Implementation
[0018] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.
[0019] like Figure 1 As shown, a pixel polarization camera is fixed above the conveyor belt. A telecentric lens is mounted on the camera, with a depth of field no less than the typical thickness of the automotive glass being tested, and precise focusing on the glass edge to ensure clear observation of both the upper and lower surfaces, as well as the ground edge area. The lens magnification ratio is matched to the camera's target size, resulting in an imaging range of approximately 30mm. This covers the unidirectional stress area of the automotive glass edge and allows for tolerances of glass positional deviations. The integrated light source is a monochromatic LED surface light source with a central aperture. A polarizer based on a polymer material is placed on its emitting surface, with a hole of the same size in the center, and is fixed in place by a glass plate. The camera lens captures images of the glass edge through the central aperture of the light source. A sensing sensor is fixed on the conveyor belt at a certain point in front of the glass. Its working principle is as follows: a beam of light is emitted from the sensor's emitting aperture. If the light is reflected onto a receiver next to the aperture, it is detected that an object has arrived, and an electrical pulse is emitted to trigger the camera to take a picture (within microseconds). The distance between the camera and the sensor is adjustable depending on the glass size, ensuring that the shooting position is near the center of the glass edge (the sampling point specified by national standards). Polarized light emitted from the integrated light source undergoes diffuse reflection on the conveyor belt after passing through the glass, but its polarization properties are largely preserved. This diffusely reflected light acts as the background light for transmission-type detection (e.g., Figure 2 As shown in the image, when the image passes through the glass again, carrying the stress information of the glass, it is captured by the camera. A pixel polarization camera is a special type of camera. Each pixel has a polarizer made of a nano-grating in front of it. Every 2x2 pixels form a group, with each pixel in the group having polarization directions of 0°, 45°, 90°, and 135°. After taking a picture, the camera extracts pixels with the same polarization direction from each group to form a sub-image. Thus, each time the camera takes a picture, it obtains four sub-images corresponding to four different polarization directions, denoted as I0, I... 45 , I 90 , I 135 These four polarization images can be used to obtain the stress-related phase difference δ (which has a linear relationship with stress; multiplying it by a coefficient converts it to stress) through the following calculation. Sin δ=[(I0-I 90 ) 2 +(I 45 -I 135 ) 2 ] 0.5 / (I0+I90 (3) Compared to direct transmission detection, the light diffused by the conveyor belt is weaker and its polarization performance is also reduced, resulting in poor contrast in the four polarized images captured by the camera. Figure 3 As shown in (a), however, due to the need for subtraction and enhancement operations between the four images during calculation, identical background portions across the four images are subtracted, while different or varying foreground information is preserved and amplified, resulting in the final image shown. Figure 3 The result shown in (b) represents information about the photoelastic fringes and their distribution. Using image recognition and artificial intelligence algorithms, the photoelastic fringes order and center position of the glass edge area and surrounding areas can be identified within 2 seconds. The maximum stress at the glass edge can then be obtained through fitting and extrapolation. From the sensor identifying the glass to the triggering camera taking a picture and completing the stress test at one detection point, the entire process takes approximately 2 seconds, less than the time interval between two pieces of glass on the conveyor belt. This allows for online stress detection of products without affecting production.
[0020] The embodiments of this application have been described above with reference to the accompanying drawings. Unless otherwise specified, the embodiments and features in the embodiments of this application can be combined with each other. This application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.
Claims
1. A method for reflective online detection of edge stress in automotive glass, characterized in that: Includes the following steps: A pixel polarization camera is fixed above the conveyor belt. The camera is equipped with a telecentric lens. The telecentric lens has a depth of field that is not less than the thickness of the glass and is precisely focused on the edge of the glass. The lens magnification ratio is adapted to the size of the camera target surface, so that the imaging range covers the unidirectional stress area of the edge of the car glass and a margin is reserved for glass position deviation. An integrated light source is set in the camera optical path. The integrated light source is a monochromatic LED surface light source with a hole in the middle. Its light-emitting surface is covered with a circular polarizer with a polymer material substrate. The polarizer has a hole in the middle with a size equivalent to the central hole of the light source and is fixed by pressing with a glass plate. The camera lens takes pictures of the photoelastic stripes on the edge of the glass through the central hole of the light source. A sensing sensor is set in front of the glass travel path of the conveyor belt. When the sensing sensor detects the arrival of the glass, it emits an electrical pulse to trigger the pixel polarization camera to take a picture. The sensing sensor determines the presence of the object by emitting and receiving light beams. The distance between the camera and the sensor is adjusted according to the size of the glass to ensure that the shooting position is located at the sampling point specified by national standards at the edge of the glass. After capturing images with a pixel polarization camera, pixels with polarization directions of 0°, 45°, 90°, and 135° are extracted from each group of 2×2 pixels to form I0 and I... 45 I 90 I 135 Four sub-images; by performing subtraction and other operations and enhancement processing on the four sub-images, stress-related foreground information can be preserved and amplified. Then, the formula is used... DoLP=|Sinδ|=[(I0-I 90 ) 2 +(I 45 -I 135 ) 2 ] 0.5 / (I0+I 90 ) (1); The DoLP fringe field in the glass edge region was calculated; Then, by using image recognition and artificial intelligence algorithms, the order and center position of the photoelastic fringes in and around the glass edge are identified, the phase difference δ at the center of the fringes is obtained, and the maximum phase difference at the glass edge is obtained by fitting and extrapolating. According to national standards, the maximum compressive stress can be obtained by multiplying it by the photoelastic fringes value of the glass material and dividing it by the glass thickness. The entire testing process takes no more than 2 seconds, which is less than the time interval between two adjacent pieces of glass on the conveyor belt.
2. The method for online reflection-based detection of edge stress in automotive glass according to claim 1, characterized in that: The integrated light source and the pixel polarization camera are integrated together and placed above the glass to be inspected. The circularly polarized light emitted by the integrated light source is transmitted through the glass and forms diffuse reflection light that retains polarization properties on the conveyor belt or black ceramic coating. This diffuse reflection light, as background light, passes through the glass again and carries stress information to be captured by the camera.
3. The method for online reflection-based detection of edge stress in automotive glass according to claim 1, characterized in that: The sensing sensor actively emits light, detects the arrival of the glass by receiving the light reflected from it, and immediately sends out an electrical pulse to trigger the camera to take a picture, with a trigger response time in the microsecond range.
4. The method for reflective online detection of edge stress in automotive glass according to claim 1, characterized in that: An LED surface light source with a central hole is used, positioned directly above the point to be measured, to ensure uniform brightness at the point.
5. The method for reflective online detection of edge stress in automotive glass according to claim 1, characterized in that: The integrated light source integrates a circular polarizer into a surface light source and fixes it to the light-emitting surface by pressing it with a glass plate.
6. The method for online reflection-based detection of edge stress in automotive glass according to claim 1, characterized in that: The camera used is a pixel polarization camera, located in the middle of the integrated light source with a hole, and images are formed on the edge of the glass through the central hole of the light source.
7. The method for online reflection-based detection of edge stress in automotive glass according to claim 1, characterized in that: Image enhancement processing is achieved by adjusting I0 and I... 90 I 45 I 135 The four sub-images are subtracted to eliminate identical background parts, while retaining and amplifying stress-related foreground variation information.
8. The method for reflective online detection of edge stress in automotive glass according to claim 1, characterized in that: When using image recognition and artificial intelligence algorithms, the identification of the photoelastic fringes level and center position of the glass edge grinding area and nearby areas is completed within 2 seconds. Combined with the DoLP field results calculated by formula (1), the compressive stress at the outermost edge of the car glass is determined. The specific calculation process is roughly as follows: Step 1: Within a 30mm radius of the edge, perform a stripe search perpendicular to the edge. If the DoLP (Dot LP) is less than 0.2 for more than 10 consecutive pixels, it is considered a black stripe, and the search continues forward to determine its width. Within the width range, the gradient of pixel values is used as the criterion, and the point with the minimum gradient is determined as the center of the black stripe (y). bi Using the same processing method, and taking 10 or more consecutive pixels with a DoLP greater than 0.5 as the criterion, find the white stripes and their centers (y). wi ); Step 2: Determine the level of the black and white stripes according to the search order: the first black stripe is level 0, the second black stripe is level 1, and so on; the first white stripe is level 0.5, the second is level 1.5, and so on. Step 3: Assign a phase difference value to the centers of the black and white stripes, with the phase difference at the center of the black stripe being δ. bi =n b *π, n b The black stripe sequence is given by a phase difference of δ between the centers of the white stripes. wi =n w *π,n w The order of the white stripes; Step 4: Construct data pairs (y-coordinates) of the center position and phase difference of the black and white stripes. i , δ i Polynomial fitting is performed and interpolation is performed until the outermost edge of the glass to obtain the phase difference at that point; Step 5: Multiply the phase difference at the outermost edge of the glass by the stripe value of the glass material to obtain the (maximum) compressive stress in the glass edge region.
9. The method for online reflection-based detection of edge stress in automotive glass according to claim 1, characterized in that: The installation position of the sensing sensor and the distance between it and the camera are dynamically adjusted according to the size of the car glass to be tested, ensuring that the shooting position is above the inspection point on the edge of the glass. The imaging lens is a telecentric lens with a depth of field no less than the thickness of the glass, and the imaging range covers an area of approximately 30mm from the edge of the glass.